DQNã‚’Kerasã¨TensorFlowã¨OpenAI Gymã§å®Ÿè£…ã™ã‚‹ã¯ã˜ã‚ã« å°‘ã—時代é…ã‚Œã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ãŒã€å¼·åŒ–å¦ç¿’ã®æ‰‹æ³•ã®ã²ã¨ã¤ã§ã‚ã‚‹DQNã‚’DeepMindã®è«–æ–‡Mnih et al., 2015, Human-level control through deep reinforcement learningã‚’å‚考ã«ã—ãªãŒã‚‰ã€Kerasã¨TensorFlowã¨OpenAI Gymを使ã£ã¦å®Ÿè£…ã—ã¾ã™ã€‚ å‰åŠã§ã¯è»½ãDQNã®ãŠã•ã‚‰ã„ã‚’ã—ã¾ã™ãŒã€å°‘ã—ã®å¼·åŒ–å¦ç¿’ã®çŸ¥è˜ã‚’æŒã£ã¦ã„ã‚‹ã“ã¨ã‚’å‰æã«ã—ã¦ã„ã¾ã™ã€‚ ã™ã§ã«ã„ãã¤ã‹è‰¯è¨˜äº‹ãŒå‡ºã¦ã„ã‚‹ã®ã§ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚åˆã‚ã›ã¦èªã‚€ã¨ç†è§£ã®åŠ©ã‘ã«ãªã‚‹ã¨æ€ã†ã®ã§ã€æ˜¯éžå‚考ã«ã—ã¦ã¿ã¦ãã ã•ã„。 DQNã®ç”Ÿã„ç«‹ã¡ã€€ï¼‹ã€€Deep Q-Networkã‚’Chainerã§æ›¸ã„㟠DQNãŒç”Ÿã¾ã‚ŒãŸèƒŒæ™¯ã«ã¤ã„ã¦èª¬æ˜Žã—ã¦ãã‚Œã¦ã„ã¾ã™ã€‚Chainerã§ã®å®Ÿè£…ã‚‚ã‚ã‚‹ãã†ã§ã™ã€‚ ゼãƒã‹ã‚‰Deepã¾ã§å¦ã¶å¼·åŒ–å¦ç¿’ タイトルã®é€šã‚Šã€ã‚¼ãƒã‹ã‚‰Deepã¾
{{#tags}}- {{label}}
{{/tags}}